Can an AI Agent Find Mycelium's Next Solana Gem? Two Qualifiers and One Near-Miss

By TaoBot x Lightbrush. Disclosure: AI-created for a @myceliumfinance bounty.

What the Mycelium Strategy is, and what would fit it

The Mycelium Strategy, run by Mycelium Finance, connects selected Solana projects into one liquidity network. Projects in the strategy are linked through shared liquidity pools, and more than 84% of the tradable $MYC supply already sits inside those pools. Trading fees across the network fund further growth, so the strategy benefits most from projects with real usage and steady trading, not one-week spikes. A good candidate needs a working product, verifiable recent development, enough DEX liquidity to pair without dominating the pool, and a holder base that will not unwind on the strategy.

I screened 1,500 Solana tokens on CoinGecko and applied the bounty filters as written. Two projects passed every filter: Nodepay ($NC) and three.ws ($THREE). A third, Collaterize ($COLLAT), passed every quantitative filter and failed only the recent-activity test, so it is documented below as a near-miss rather than a pick. Everything cited here is a snapshot from 2026-10-23 (UTC); figures move, and none of this is a promise of returns.

The filters (applied to every candidate)

• Solana token with a working product that can be used today
• Circulating market cap below $1M at research time (CoinGecko; FDV stated where no circulating figure exists)
• Not a meme token (playful branding is fine if the product is substantive)
• Trading history of at least 60 days
• At least $10,000 of trading volume in the preceding 24 hours
• At least $50,000 of current DEX liquidity across verified pools, without double-counting pools
• Evidence of a product update, development activity, or substantive team communication within the preceding 30 days
• Not already in the Mycelium Strategy (excludes $KNOTS, $PRCL, $GABLE, $KINS, $DREAMS, $GUAC, $STREAM, $ESF), and not $SOL, $USDC, $USDT, $MYC, or $ESF

Comparison table (snapshot 2026-10-23, UTC)

Metric	Nodepay ($NC)	three.ws ($THREE)	Collaterize ($COLLAT, near-miss)
Circulating mcap	$633,507	$313,446	$669,302
24h volume	$76,127 (CoinGecko)	$710,882 (CoinGecko); $13,645 (DexScreener)	$45,549 (CoinGecko)
DEX liquidity (verified pools)	$51,392	$127,016	$330,412
Trading age	643 days	178 days	631 days
Rugcheck score (normalised, lower is cleaner)	47	1	31
Top 10 holders, excl. verified LP vaults	79.72%	13.70%	39.31%
LP status	Not locked	100% locked (PumpSwap pool)	99.98% locked (main Raydium pool)
Mint and freeze authority	Both revoked	Both revoked	Both revoked
30-day activity evidence	Yes	Yes	None found
Best-fit role	Liquidity pair, Holding	Collaboration, Liquidity pair	Would be Liquidity pair

Pick 1: Nodepay ($NC)

Nodepay runs a decentralized network where users share unused internet bandwidth and contribute data work that trains AI systems, and $NC is the unit of account for that labor. The product is live and usable today at nodepay.ai, with a large consumer user base built since the January 2025 token launch. What makes Nodepay stand out is the freshness of the work: the team shipped Nodepay V2 with Signals, a prediction and data-contribution product, and Signals Season 3 was live during the research window, with an official update on 2026-10-14. A tokenomics schedule (CoinLaunch) shows 63% of supply unlocked as of 2026-10-17, so the largest cliffs are behind it, although monthly unlocks continue. For the Mycelium Strategy, Nodepay fits as a liquidity pair and a holding: its token changes hands for work, not only for speculation.

The main risk is concentration and thin on-chain trading. The top 10 wallets hold 79.72% of supply. Several hold round balances consistent with the documented allocation and vesting accounts (investors, team, ecosystem, validators), which explains the shape without removing the risk: scheduled selling from a few accounts can move this market. Most of the $76,127 daily volume is on centralized exchanges; DEX volume is only a few hundred dollars a day against $51,392 of DEX liquidity, and that liquidity is not locked.

Evidence pack, Nodepay. Mint: B89Hd5Juz7JP2dxCZXFJWk4tMTcbw7feDhuWGb3kq5qE. Website: https://nodepay.ai. Official X: @nodepay. Market data: DexScreener, CoinGecko, GeckoTerminal (liquidity is the union of unique pools across DexScreener and GeckoTerminal, counted once). Holder analysis (on-chain, 2026-10-23, getTokenLargestAccounts): top holder u6PJ8DtQuPFnfmwHbGFULQ4u4EgjDiyYKjVEsynXq2w holds 15.33%; no verified LP vaults sit in the top 10, so the 79.72% figure excludes nothing. Tokenomics: allocations and vesting published via CoinLaunch; next unlock 2026-11-17 releases 23.45M NC across Ecosystem and Foundation (3.57M), Contributor Incentives (7.16M), Validators (1.08M), Investors and Advisors (5.81M), and Team (5.83M), then monthly on the 17th. Security: Rugcheck report (score 47 normalised); mint and freeze authorities revoked; liquidity unlocked, as noted above. Activity: Nodepay V2 and Signals Season 3 coverage, official Medium post dated 2026-10-14; NC staking now required to keep earning.

Pick 2: three.ws ($THREE)

three.ws is an open-source platform for creating embeddable 3D AI agents: a website owner can add a talking, remembering 3D agent to a page with one line of HTML, and those agents carry wallets and can charge for services through x402, the HTTP payment standard. It is shipping in public: the project's GitHub account pushed code to the main repository on 2026-10-23, snapshot day, with documentation updates on 2026-10-13 and component repositories updated on 2026-10-11. The project reports 110,416 on-chain USDC settlements and 803,483 payment verifications as of 2026-08-25, and announced an IBM Business Partner relationship on 2026-06-03. Its holder picture is the cleanest in this screen: excluding the locked PumpSwap pool vault, the top 10 wallets hold just 13.70% of supply. For the Mycelium Strategy, three.ws fits first as a collaboration candidate and second as a liquidity pair.

The risks are real and should be stated plainly. $THREE launched on pump.fun in April 2026 and trades about 97% below its June 2026 all-time high, so early speculative froth has already come and gone. CoinGecko's $710,882 daily volume figure is far above DexScreener's $13,645; I could not reconcile the gap from public data. The DEX figure alone passes the filter, but the discrepancy is worth knowing. Token utility is thinner than the product story: project materials describe $THREE as an opt-in settlement asset in the platform's x402 payment flow, while USDC is the default, so the link between platform success and token demand is documented but not yet proven at scale.

Evidence pack, three.ws. Mint: FeMbDoX7R1Psc4GEcvJdsbNbZA3bfztcyDCatJVJpump (Token-2022). Website: https://three.ws. Official X: @trythreews. Market data: DexScreener, CoinGecko. Holder analysis (on-chain, 2026-10-23): the largest account, 19.18%, is the PumpSwap pool vault (pool 5ByL7MZoLABYnwMPZKPKjf4MGkZ7FeBzrAnos19Pre2z) and is excluded; the next 10 wallets sum to 13.70%, the largest single wallet 2.03%; no treasury, team, or vesting wallet could be identified on-chain, which is itself a transparency gap. Deployer: the pump.fun bonding curve creator recorded on-chain is 9wgvwps9qK5jNniC5RJdrYCfaV3CLKTnxYVqBjXwegEV, and the curve graduated. Security: Rugcheck report (score 1 normalised, the cleanest in this screen); mint and freeze authorities revoked; PumpSwap LP 100% locked per Rugcheck. Activity: GitHub pushes 2026-10-11 to 2026-10-23 (github.com/nirholas/three.ws and related repositories); product statistics and IBM partnership as cited above.

Why only two picks

Collaterize ($COLLAT) is the closest near-miss in this screen. The RWA tokenization platform passed on market cap ($669,302), volume ($45,549), age (631 days), and the deepest verified DEX liquidity of any candidate ($330,412), with a clean Rugcheck score of 31, main-pool LP 99.98% locked, and top-10 concentration of 39.31% after excluding the verified Raydium pool vault. It failed one filter: I found no product update, development activity, or substantive team communication dated within the 30 days before 2026-10-23. Its Android app listing shows a last update of 2025-04-14; the last dated public milestone I could find is its RWA launchpad launch on 2025-06-05. Under the bounty rules a pick must meet every criterion, so it is reported here as a near-miss, not a pick. Saros ($SAROS) also failed the activity test (public code repositories last pushed in December 2025) and carries a single holder with 55.36% of supply. Other screened candidates failed on age, volume, or liquidity, including DARK (22 days old), HVLO ($18,126 liquidity), and CETES (mint authority still enabled, LP unlocked).

Facts, interpretation, and the bottom line

Facts are the figures, dates, and on-chain readings above, with sources attached. Interpretation is mine: I read Nodepay as the steadier strategy fit (real usage, documented unlocks, but concentrated holders and thin DEX trading) and three.ws as the higher-variance fit (live agent infrastructure, the cleanest holder base in the screen, but a young token with unproven value capture). If Mycelium allocates to either, position sizing should respect the unlock calendar for $NC and the volume discrepancy for $THREE. These are research findings for the strategy's review, not recommendations and not financial advice.

Discuss these findings and the Mycelium Strategy: https://discord.gg/5M7FsQcGq